Download On the origin of noble gases in mantle plumes

10.1098/ rsta.2002.1084
On the origin of noble gases in mantle plumes
By N ic o l a s Co ltic e1 a n d Y an ick R i c a r d2
1
Department of Geosciences, Guyot Hall, Princeton University,
Princeton, NJ 08544-1003, USA ([email protected])
2 Laboratoire de Sciences de la Terre, Ecole Normale Sup¶erieure de Lyon,
46 all¶ee d’Italie, 69364 Lyon cedex 07, France ([email protected])
Published online 23 September 2002
The chemical di¬erences between deep- and shallow-mantle sources of oceanic basalts
provide evidence that several distinct components coexist within the Earth’s mantle.
Most of these components have been identi­ ed as recycled in origin. However, the
noble-gas signature is still a matter of debate and questions the preservation of
primitive regions in the convective mantle. We show that a model where the noblegas signature observed in Hawaii and Iceland comes from a pristine homogeneous
deep layer would imply a primitive 3 He content and 3 He/22 Ne ratio that are very
unlikely. On the contrary, mass balances show that the partly degassed peridotite of
a marble-cake mantle can be the noble-gas end-member with an apparent `primitive’like composition. This component is mixed with recycled oceanic crust in di¬erent
proportions in the plume sources and in the shallow mantle. A recycling model of
the mantle, involving gravitational segregation of the oceanic crust at the bottom of
the mantle, potentially satis­ es trace-element as well as noble-gas constraints.
Keywords: mantle; noble gas; plumes; recycling; convection
1. Introduction
Isotope geochemistry of mantle-derived rocks provides fundamental information
about the mantle’s chemical structure and dynamics. Systematic chemical di¬erences between mid-ocean ridge basalts (MORBs) and oceanic basalts (OIBs) are
observed and show that chemically di¬erent domains coexist within the Earth’s
mantle (Zindler & Hart 1986; Hofmann 1997). Global mass balances have shown
that signi­ cant portions of the mantle have to be richer in trace elements than in the
shallow mantle (Jacobsen & Wasserburg 1979; Davies 1981). Although the nature of
these regions cannot be determined from this approach, further insights are expected
from isotopic studies of mantle-derived basalts.
The identi­ cation of the isotopic end-members is a direct interpretation of the
isotopic ratios measured in oceanic basalts. Isotopic systems like 18 O{16 O, U{Pb,
Lu{Hf and Re{Os have been used as ­ ngerprints to identify recycled crustal components in the sources of Polynesia (Hauri & Hart 1993), Hawaii (Eiler et al. 1996;
Blichert-Toft et al . 1999a) and Iceland (Chauvel & H´ emond 2000) among others. The
di¯ culties in understanding the isotopic signal of oceanic basalts come mostly from
One contribution of 14 to a Discussion Meeting `Chemical reservoirs and convection in the Earth’s
mantle’.
Phil. Trans. R. Soc. Lond. A (2002) 360, 2633{2648
2633
°
c 2002 The Royal Society
N. Coltice and Y. Ricard
2634
noble gases. Some hotspots, like Loihi and Iceland, have signi­ cantly lower 4 He/3 He
and 21 Ne/22 Ne ratios than MORBs (Kurz et al . 1982; Sarda et al . 1988; Dixon et al .
2000). Since 3 He and 22 Ne are primordial isotopes, contrary to the radiogenic 4 He
and nucleogenic 21 Ne, it is commonly assumed that Loihi and Iceland are derived
from pristine, undegassed portions of the mantle. In addition, the 40 Ar{40 K budget
of the mantle tends to favour the existence of an undegassed reservoir in the lower
mantle (All³egre et al. 1996).
The understanding of the geochemical signal is closely related to the competition
between recycling degassed products and preserving undegassed regions in vigorous
convection. The most recent geophysical observations show clear evidence of deep
subduction. Global tomography over North America indicates that the Farallon plate
is sinking through the whole mantle (Grand 1987; van der Hilst et al . 1997). This
mode of convection is also able to explain more than 80% of geoid anomalies (Ricard
et al . 1993). In a range of Earth-like parameters, convection calculations predict a
deep-mantle ®ow and reproduce many of the observables (Bunge et al . 1998). In
this context, the preservation of large-scale primitive regions is very unlikely, as is
suggested by several mixing studies (van Keken & Ballentine 1998; Ferrachat &
Ricard 2001).
Strictly speaking, mantle convection implies stirring, i.e. stretching and folding,
not mixing, since chemical di¬usion is negligible. The survival of highly deformed
primitive veins in the mantle is likely. Even in the case of random distribution of
subducted material in the mantle, the proportion of mantle rocks that have never
been processed at ridges would never be zero but would decrease exponentially with
a residence time F =M comparable with the age of the Earth (where M is the mantle
mass and F the mass ®ux of subducted material). Stirring is certainly more complex
than simple random redistribution, but all numerical simulations agree that a small
but signi­ cant volume of material (from 10 to 35%) has not seen the surface (van
Keken & Ballentine 1998; Ferrachat & Ricard 2001; Coltice et al . 2000b).
Keeping any sort of chemical layering implies the existence of a stabilizing density
strati­ cation, and preserving an undegassed abyssal layer in the deep mantle requires
that primitive rocks are slightly denser than the overlying di¬erentiated material
(Kellogg et al . 1999). Another point of view is that chemical layering is an ongoing
process as a consequence of plate tectonics: the dense eclogitic oceanic crust can
accumulate at the bottom of the mantle by gravitational segregation (Christensen
& Hofmann 1994). This model could also explain various geochemical observations
(Coltice & Ricard 1999; Ferrachat & Ricard 2001). In this paper, we attempt to identify the origin of the `primitive’-like noble-gas signature observed in several hotspots,
in order to evaluate various mantle models.
2. Noble-gas constraints on mantle heterogeneity
(a) Isotopic ratios
Noble-gas systematics of oceanic basalts show signi­ cant di¬erences between the
shallow-mantle source of MORBs and the sources of several OIBs, particularly Loihi
(Kurz et al . 1982; Valbracht et al. 1997; Trielo¬ et al . 2000) and Iceland (Dixon et al .
2000; Trielo¬ et al . 2000; Harrison et al . 1999). These di¬erences are identi­ ed by the
ratios of radiogenic or nucleogenic isotopes over primordial isotopes, mostly 4 He/3 He,
21 Ne/22 Ne and 40 Ar/36 Ar. 4 He and 21 Ne are produced through the radioactive decay
Phil. Trans. R. Soc. Lond. A (2002)
On the origin of noble gases in mantle plumes
14
(a)
2635
SOLAR
20Ne/22Ne
13
12
11
10
AIR
9
0.03
0.04
0.05
0.06
0.07
0.08
21Ne/22Ne
(b)
90 000
80 000
4He/3He
70 000
60 000
50 000
40 000
30 000
0
5000
10 000 15 000 20 000 25 000 30 000
40Ar/36Ar
Figure 1. Neon isotope plot and argon{helium isotope plot for the popping rock (open circles
(Moreira et al . 1998)), Iceland samples (¯lled circles (Dixon et al . 2000; Trielo® et al. 2000;
Harrison et al . 1999)) and Loihi samples (open squares (Valbracht et al . 1997; Trielo® et al .
2000)).
of 235;238 U and 232 Th, whereas 40 Ar is produced from 40 K by electron capture. A
fundamental problem in noble-gas geochemistry is that the initial isotopic signature
of a basalt is often perturbed by the addition of atmospheric gases (except for helium,
which is lost into space out of the atmosphere) (Ballentine & Barfod 2000). Correcting
for air contamination can be performed using the ratio of two stable isotopes of neon,
20 Ne/22 Ne (Farley & Poreda 1993), which is presumed to be solar or maybe meteoritic
(Trielo¬ et al . 2000).
When corrected from atmospheric contamination, MORBs show systematically
higher radiogenic/primordial ratios compared with Loihi and Iceland, as depicted in
­ gure 1. These di¬erences are attributed to lower 3 He/(Th+U), 22 Ne/(Th+U) and
36 Ar/K in the shallow mantle. The standard model argues for a primitive component
Phil. Trans. R. Soc. Lond. A (2002)
N. Coltice and Y. Ricard
2636
Table 1. Typical isotopic composition of the shallow mantle and `primitive’-like plumes
(Shallow-mantle data from Moreira et al . (1998), Loihi data from Trielo® et al . (2000) and
Valbracht et al . (1997) and Iceland data from Dixon et al . (2000), Trielo® et al . (2000) and
Harrison et al . (1999).)
4
shallow mantle
Loihi
Iceland
He/3 He
80 000{90 000
24 000{40 000
20 000{40 000
40
Ar/ 36 Ar
30 000{45 000
2 500{8 500
3 000{8 000
21
Ne/22 Ne
>0.07
0.03{0.04
0.03{0.04
in the source of plumes, rich in primordial noble gases (Porcelli & Wasserburg 1995).
That assumption will be discussed below. Whatever the model, at least two components have to coexist within the mantle and they must have signi­ cant di¬erences
in ratios between primordial rare gases and radiogenic parent.
(b) Helium concentrations
4 He
The
abundance in the shallow mantle can be deduced from the 4 He degassing
®ux from the mantle (Jean-Baptiste 1992; Farley et al . 1995). This ®ux is measured with a large uncertainty, ca. 9 £ 107 mol yr¡1 . The emplacement rate of oceanic
crust is close to 6 £ 1016 g yr¡1 . Consequently, the 4 He content of the undegassed
magma should be 1500 £ 10¡12 mol g¡1 . Assuming that helium is incompatible and
that the newly formed oceanic crust comes from ca. 10% partial melting gives a
4 He content of ca. 150 £ 10 ¡12 mol g¡1 . The uncertainty on this value is at least a
factor of two, but we consider for numerical applications that the upper bound is
ca. 300 £ 10¡12 mol g¡1 . A larger value would not invalidate our conclusions.
The 4 He content in the source of Loihi cannot be computed directly; however, it
has been noted that OIBs have low helium concentrations with respect to MORBs.
This observation is referred to as the `helium paradox’, since OIBs are thought to
derive from less-degassed sources (Anderson 1998). Can some kind of helium loss
rule out this paradox? One way to estimate the helium loss is to compare the 3 He/U
ratio of OIBs and MORBs (the He/U ratio in the magma is identical to that in the
source, as they are both very incompatible). According to their low 4 He/3 He ratio,
OIBs should have higher 3 He/U ratios than MORBs. A typical uranium-content
measurement in Loihi is three to four times higher than that in a fresh MORB (Sims
& De Paolo 1997), so the helium content of a Loihi magma prior to degassing has
to be, at least, three to four times that of MORBs. This indicates that most of the
original helium is degassed before the magma reaches the surface. Hence, part of the
`helium paradox’ is due to near-surface degassing.
(c) 3 He/ 22 Ne
Another constraint comes from the measured elemental composition. One must
keep in mind that the elemental composition of samples can be fractionated by partial
melting and degassing processes (Moreira & Sarda 2000). However, it is possible
to estimate 3 He/22 Ne in the di¬erent mantle sources using the production ratio
21 Ne/4 He of 4:5 £ 10 ¡8 (Yatsevich & Honda 1997) (both 21 Ne and 4 He are produced
Phil. Trans. R. Soc. Lond. A (2002)
On the origin of noble gases in mantle plumes
2637
by the decay of U and Th),
3 He
22 Ne
=
3 He 21 Ne 4 He
4 He 22 Ne 21 Ne
;
(2.1)
where 21 Ne/22 Ne is corrected for air contamination. Using the values of table 1,
3 He/22 Ne is around 13 for MORBs and for both Loihi and Iceland samples (Harrison
et al. 1999). However the 3 He/22 Ne measured directly on a popping rock (i.e. a highly
vesiculated basalt) (Moreira et al. 1998) and on Loihi glasses (Valbracht et al . 1997)
is around seven. But whatever the preferred values, the di¬erent studies lead to
the same idea: both `primitive’-like OIB and MORB sources have roughly similar
3
He/22 Ne (Honda & Patterson 1999). The elemental ratio 3 He/36 Ar is signi­ cantly
more di¯ cult to retrieve because of the di¯ culties in correcting for atmospheric
36 Ar and because the production ratio 4 He/40 Ar is a function of the age of the
closed system (unlike 21 Ne and 4 He, 40 Ar and 4 He do not belong to the same decay
chain).
The relative constancy of 3 He/22 Ne among oceanic basalts is surprising, as He
and Ne are fractionated in the residue after partial melting. This observation means
either that the shallow and deep mantles have undergone the same degassing history
or that a similar gas-rich component is present in both regions.
3. A primitive layer as the origin of plume noble gases?
A primitive homogeneous layer supplying noble gases in mantle plumes is a frequent
hypothesis. However, this section shows some of the limitations and inherent contradictions of this model.
(a) Primitive fraction in mantle plumes
In the model involving the persistence of pristine regions in the mantle, the
`primitive’-like noble-gas signature of some OIBs (`oib’) derives from a mixing
between the shallow mantle (SM) and a mass fraction f of a primitive mantle (PM)
source (see ­ gure 2). For simplicity we neglect the recycling components that are
also present in OIBs.
For example, the mass-balance equations for any isotope concentration [X ] reads:
[X ]oib = f [X ]PM
+ (1 ¡
f )[X ]S
M
:
(3.1)
In this equation, X can be substituted by either R (a radiogenic isotope like 40 Ar or
4 He) or P (a primitive isotope like 36 Ar or 3 He). With a little algebra, we get
[R]S
M
= [R]PM
f
1¡
(P=R)PM ¡ (P =R)oib
:
f (P=R)oib ¡ (P =R)S M
(3.2)
Applying the previous equation to the 40 Ar{36 Ar system can provide a fair estimate
of f because the mantle{atmosphere system is closed for argon, and all the terrestrial
40 Ar comes from the decay of 40 K. The primitive 40 Ar/36 Ar of the mantle cannot
exceed 400, assuming that the shallow mantle, atmosphere and continental crust are
complementary ((P=R)PM ¹ 1=400) (Turner 1989). The 40 Ar/36 Ar of OIB and SM
are chosen from the measured values of 2500 (lower bound for Loihi, (P =R)oib ¹
1=2500) and 40 000 ((P =R)S M ¹ 1=40 000), respectively (see table 1).
Phil. Trans. R. Soc. Lond. A (2002)
N. Coltice and Y. Ricard
2638
(a)
lithosphere
shallow mantle
plume containing
pristine mantle
(b)
well-mixed
shallow mantle
poorly mixed
plume
Figure 2. Schematic sketches of alternative models for `primitive’ -like plume noble gases.
(a) Noble gases of OIBs are explained by the presence of pristine material in plumes. (b) Variable
mixings of oceanic crust and residual peridotite provide the noble-gas variety of oceanic basalts.
A present-day [40 Ar]PM of ca. 940 £ 10¡12 mol g¡1 can be estimated from a primitive potassium content of 240 ppm (McDonough & Sun 1995). [40 Ar]S M is di¯ cult
to measure accurately. However, a range of acceptable values for the 40 Ar concentrations can be determined from the 4 He degassing ®ux discussed earlier. Taking the
lower bound for the 4 He/40 Ar measured in a popping rock to be 1.5 provides an
upper bound for the shallow mantle 40 Ar content of 200 £ 10¡12 mol g¡1 (Moreira et
al . 1998). This value implies a 40 Ar degassing ®ux of 1:2 £ 107 mol yr¡1 in agreement
with observations (Stuart & Grenville 1998).
All these concentration and ratio estimates used in equation (3.2) lead to an upper
bound value of f of only 3% in the source of Loihi basalts. The conclusion is therefore
[40 Ar]oib ¹
36
[ Ar]oib ¹
[40 Ar]S
M
;
36
f [ Ar]PM :
[36 Ar]
(3.3)
(3.4)
The primitive mantle is so rich in primitive
that even a minute fraction of
it su¯ ces to give Loihi its primitive ®avour. This is not the case for the radiogenic
[40 Ar], which is not a¬ected by a few per cent of primitive material. This is not the
case for common trace elements either. For example, 3% of primitive mantle having
a uranium content of 21 ppb, will only contribute to 5% of the uranium content of
the plume source, assuming 8 ppb of uranium in the shallow mantle (Zindler & Hart
Phil. Trans. R. Soc. Lond. A (2002)
On the origin of noble gases in mantle plumes
2639
1986). In the hypothesis of a primitive layer supplying noble gases, the composition
and isotope ratios of incompatible refractory elements in the sources of Loihi or
Iceland should be similar to those of the source of MORBs.
(b) Elemental composition of the primitive material
Now that we know that the fraction of primitive material in plumes is small
(f ¹ 3%), we can estimate the 3 He content in the primitive material. Reformulating
equation (3.2), we get
µ ¶
½µ ¶
µ ¶ ¾
P
1¡ f
P
P
[P ]PM = [R]PM
+
[R]S M
¡
;
(3.5)
R oib
f
R oib
R S M
where we use P = 3 He and R = 4 He.
A source of uncertainty in the numerical estimate comes from [4 He]PM because
an unconstrained quantity of 4 He must have been trapped during Earth’s accretion.
But the closed-system production ratio 4 He/40 Ar of 1.5 gives a lower bound for
[4 He]PM . The other quantities have already been discussed. The obtained primitive
3 He is then a lower bound, ca. 0:2 £ 10 ¡12 mol g¡1 , slightly higher than the value
computed by Harper & Jacobsen (1996) from a comparable analysis. This value is
as high as the concentration of typical carbonaceous chondrites (Mazor et al . 1970).
However, shock experiments suggest that more than 99% of the initial volatile content
of the parent bodies should have been degassed during Earth’s accretion (Tyburczy
et al. 1986). For that reason, it has been proposed that such a high helium content
has been trapped from a solar composition proto-atmosphere (Harper & Jacobsen
1996) or from an He-rich reservoir such as interplanetary dust particles (Pepin et al .
2001). In any case, the primitive mantle should have been very rich in rare gases and
the extensive degassing of the upper mantle (99%) should not have fractionated the
residual helium and neon.
4. A degassed and ubiquitous source for `primitive’-like noble gases
(a) Noble gases in the marble-cake mantle
As the hypothesis of a primitive deep mantle seems to lead to various contradictions, the `pristine’-like noble-gas isotopic ratios are probably derived from a source
that has experienced some degree of degassing. There are two ubiquitous components in the mantle: partly degassed peridotite and stirred ancient oceanic crust (the
partly degassed peridotite being itself a mixture of stirred lithospheric mantle and of
primordial mantle). These components have been identi­ ed petrologically and geochemically in peridotite massifs and MORBs (Pearson et al . 1991; Blichert-Toft et
al . 1999b; Eiler et al. 2000). The convective mantle, made from pyroxenite layers
embedded in a matrix of peridotite, looks like a marble cake (All³egre & Turcotte
1986).
In the marble-cake mantle, the radiogenic noble gases are concentrated in the
stirred oceanic crust, rich in uranium, thorium and potassium. The crust has lost
most of its primordial volatiles and, consequently, the peridotite should be the reservoir of primordial noble gases. These two components therefore have di¬erent noblegas signatures and mixing them in variable proportions could explain the heterogenePhil. Trans. R. Soc. Lond. A (2002)
N. Coltice and Y. Ricard
2640
OIBs
4He/3He
SM
Tf
atm
L(b)
M
L(a)
OC
Ti
U/3He
Figure 3. Schematic of the evolution of the various components of the marble-cake mantle.
Between melting events, these components evolve towards lower U/3 He. However, due to the
° ux of helium to the atmosphere, their mixture evolves towards larger U/3 He.
ity observed among OIBs. There is a wealth of evidence suggesting the presence of
unusual amounts of recycled components in plume sources (e.g. St Helena, Tristan,
etc.), which have, as expected, radiogenic noble-gas ratios (Kurz et al. 1982; Hanyu
& Kaneoka 1997). On the contrary, Iceland or Loihi basalts have unradiogenic noblegas isotope ratios, which may be the ­ ngerprints of an excess of peridotite in their
sources (­ gure 2).
The evolution of the rare-gas isotopic compositions in the marble-cake mantle
can be understood using a U/3 He versus 4 He/3 He plot. In ­ gure 3, we consider the
evolution of a mantle (M) that undergoes a single event of processing at ridges.
During melting, the U/3 He is fractionated in the atmosphere (Atm), the mantle
lithosphere (L) and the oceanic crust (OC). According to the relative incompatibilities of U and He, the lithosphere can plot on the right (a) or on the left (b) of
the initial mantle composition. After some time and before a new episode of partial melting, these di¬erent reservoirs are aligned on the isochron Tf . The mantle
at time Tf is a mixture of the recycled components L and OC with some nonprocessed component M. The 4 He/3 He of the mantle spans the domain from the
most primitive, L(b) or M, to the most radiogenic, OC. This domain could be sampled by hotspots, while the MORB source would correspond to a more e¯ cient
mixing of these di¬erent components. The shallow-mantle composition at time Tf
has of course to have a higher U/3 He ratio than that at time Ti , (M). Of course,
­ gure 3 only depicts one event of melting transport, while many have occurred in
the Earth.
(b) Noble gases in the peridotite
In contrast to the classical model, where some OIBs were a mixture between primitive mantle and shallow mantle, it is now not only the shallow mantle but also the
plumes that appear as a mixture of variable quantities of oceanic crust (`crust’) and
peridotite (`per’). For example, equations (3.1) and (3.2) are now applied to helium
Phil. Trans. R. Soc. Lond. A (2002)
On the origin of noble gases in mantle plumes
2641
[4He]SM (10-12 mol g-1)
300
200
100
0
0.05
0.1
fraction of oceanic crust
0.15
Figure 4. 4 He concentration in the shallow mantle modelled as a function of the fraction of
recycled oceanic crust in a marble-cake mantle. Di® erent accumulation times of 4 He in the
oceanic crust are represented: ||, 500 Myr; ¢ ¢ ¢ ¢ ¢ ¢ ¢, 1000 Myr; { { {, 1500 Myr; { ¢ { ¢ {,
2000 Myr. The observed shallow mantle 4 He content ranges from 100 to 300 £ 10¡ 12 mol g¡ 1 .
in the shallow mantle:
[4 He]S
µ3 ¶
He
4 He
S
= (1 ¡
M
M
= (1 ¡
fc )[4 He]p
+ fc [4 He]cru s t ;
µ
¶
[4 He]p er 3 He
[4 He]cru
fc ) 4
+
f
c
4 He
[ He]S M
[4 He]S M
p er
(4.1)
er
s t
µ3
He
4 He
¶
:
(4.2)
cru s t
The concentration of radiogenic nuclides in the recycled oceanic crust is a function of its radioactive-element content (70 ppb of uranium for typical fresh MORBs
(Hofmann 1988)) and its age. The 4 He/3 He in the ancient oceanic crust should be
very high (ca. 107 ) compared with MORBs, since it has lost most of its primordial rare gases by shallow degassing. Therefore, the peridotite has to appear more
primordial than the MORB source. For simplicity, the peridotite is assumed to be
the `primitive’-like isotopic end-member in OIBs in this model (i.e. (4 He/3 He)p er =
(4 He/3 He)oib = 20 000). Three unknowns are dominant in this problem: fc , the fraction of recycled crust in the shallow mantle, the age of this crust and the 4 He content
of the shallow mantle. The relationships between these variables are shown in ­ gure 4.
For the previously estimated range of shallow mantle 4 He abundance, the modelled
age of the crust is likely to lie between 1 and 2 billion years, corresponding to a value
of fc greater than 3%. This amount of recycled crust would provide at least 2 ppb of
uranium in the MORB source, so that the peridotite would have a concentration of
uranium of a few parts per billion.
How do these predictions compare with ­ rst-order observations? The 4 He/U ratio
of mantle melts can be used to determine an approximate accumulation time of 4 He of
the shallow-mantle source, especially because both uranium and helium are supposed
to be very incompatible elements and should not be fractionated in the melts. The
Phil. Trans. R. Soc. Lond. A (2002)
N. Coltice and Y. Ricard
2642
previous estimates for the ®ux of 4 He and U content in the oceanic crust lead to an
accumulation time of more than 1.7 billion years for the mid-ocean-ridge source rocks,
with an uncertainty of 50%. The same result can be reached by calculating the 4 He
accumulation time for the popping rock ([4 He] ¹ 390 £ 10¡12 mol g¡1 ), assuming a
uranium content comparable with 70 ppb, typical of fresh MORBs. The fraction of
crust within the shallow mantle is di¯ cult to estimate, but most of the geological
studies of peridotitic massifs and MORB chemistry provide estimates ranging from
2 to 10% (Pearson et al . 1991; Blichert-Toft et al . 1999b; Eiler et al . 2000; All³egre
& Turcotte 1986). The observations concerning the `age’ and the fraction of recycled
crust of the shallow mantle con­ rm that, to ­ rst order, a partly degassed peridotite
can account for `primitive’-like noble-gas ratios.
The non-radiogenic noble gases in the shallow mantle mostly reside in the peridotite. The recycled crust is strongly degassed and hence equation (4.2) is well
approximated by
[3 He]S M ¹ (1 ¡ fc )[3 He]p er ;
(4.3)
and this is certainly true for most of the stable rare-gas isotopes. Hence, the model
predicts that the amounts of 3 He, 22 Ne and 36 Ar are the same in the shallow mantle
and the `primitive’-like plume source. We consider that this prediction is consistent
with observation (Honda & Patterson 1999; Moreira et al. 2001).
The present-day 3 He content of the shallow mantle, deduced from degassing
observations, is close to 1:5 £ 10¡15 mol g¡1 and hence approximately 100 times
smaller than the chondritic value. Several numerical simulations show that convection
dynamics, incorporating melting, degasses no more than ca. 65% of the primordial
3 He (van Keken & Ballentine 1998; Ferrachat & Ricard 2001). Hence, the prediction
of the 3 He content by the marble-cake mantle model is consistent with whole-mantle
thermochemical calculations, using a 98.5% degassed chondrite as a proxy for the
primitive Earth, i.e. it does not require to start from a gas-rich mantle.
The proposed origin of `primitive’-like noble gases seems to be consistent with
the present-day rare-gas data of MORBs and OIBs. The most-simpli­ ed version of
the model has been described here. The di¬erence in age between OIB and MORB
sources, and the impact of possible atmospheric volatile recycling (Sarda et al . 1999)
could potentially change some details.
5. The argon constraint: a deep reservoir of radiogenic gases
If the isotopic heterogeneity is the most controversial aspect of the geochemistry
of mantle noble gases, the global budget of 40 Ar is certainly a crucial issue as
well. The 40 Ar economy of the Earth potentially gives information on the degree
of degassing of mantle reservoirs. Previous studies argue that the atmosphere and
the upper mantle only combine for 60% of the global budget of 40 Ar, leaving the
remaining 40% to an undegassed reservoir in the deep mantle (All³egre et al . 1996).
Some knowledge of the potassium content in the primordial mantle is required to
reach this conclusion. The proposed potassium concentration is determined by scaling to the better-established uranium concentration, assuming that potassium and
uranium never fractionate among physical and chemical processes (Jochum et al .
1983). Albar³ ede (1998) pointed out that this assumption could be falsi­ ed and a
lower primitive K/U would question the robustness of the 40 Ar budget constraint on
Phil. Trans. R. Soc. Lond. A (2002)
On the origin of noble gases in mantle plumes
2643
2
(1016 kg)
1
excess
0
40Ar xs
DM
deficit
-1
-2
-3
1022
1023
MDM (kg)
1024
1025
40
Figure 5. Ar excess of a deep-mantle reservoir predicted, taking into account the uncertainties
in potassium and argon concentrations in the shallow mantle: ||, shallow-mantle concentrations of 80 ppm of potassium and 50 £ 10¡ 12 mol g¡ 1 of 40 Ar; ¢¢¢¢¢, shallow-mantle concentrations
of 50 ppm and 100 £ 10¡ 12 mol g¡ 1 ; { { {, the hypothesis of closed-system shallow mantle.
mantle structure. His arguments are well illustrated by the variability of K/U among
oceanic basalts (Halliday et al. 1995).
The 40 Ar budget of the silicate Earth is given by
(40 KCC +
40
KS
M
+
40
KD
M
)R(e¶
T
¡ 1) =
40
ArAtm
+ CC
+
40
ArS
M
+
40
ArD
M
; (5.1)
where 40 K and 40 Ar denote the masses of potassium-40 and argon-40. The di¬erent
reservoirs considered are the atmosphere (Atm), the continental crust (CC), the
shallow mantle (SM) and the deep mantle (DM). R is the branching ratio, ¶ is the
radioactive decay constant of potassium-40 and T is the age of the Earth. However,
the 40 Ar budget can be evaluated without making any strong assumption on the
K content of the primitive mantle. To remove the potassium concentrations from
equation (5.1), one can use the calculation of excess argon introduced by Coltice et
al . (2000a) as
40
Arxs = 40 Ar ¡ 40 KR(e¶ T ¡ 1):
(5.2)
The excess argon expresses the degree of decoupling between potassium and argon in
a reservoir, which is essentially due to degassing of K-bearing rocks. A negative 40 Arxs
is associated with a degassed reservoir, a zero value corresponds to a closed system
and a positive value is associated with a gas-rich (or potassium poor) reservoir, such
as the atmosphere. Equation (5.1) then becomes
40
ArxsAtm
+ CC
+
40
ArxsSM
+
40
ArxsD M
=0
(5.3)
and expresses the balance of the excess argon at the surface of the Earth (atmosphere plus continental crust) and the de­ cit argon in the silicate mantle (divided in
distinctive shallow and deep regions).
Phil. Trans. R. Soc. Lond. A (2002)
N. Coltice and Y. Ricard
2644
The amount of excess argon in the crust{atmosphere system is close to 1:9£1016 kg
(see Coltice et al . 2000a). The de­ cit of 40 Ar in the shallow mantle cannot be determined accurately, since its 40 Ar and K contents are not well documented. The estimate used previously, based on the popping rock, suggests that the shallow-mantle
concentration of 40 Ar is 50{200 £ 10¡12 mol g¡1 . The shallow-mantle concentration
of potassium probably lies between 30 and 80 ppm (Jochum et al . 1983). Varying
these parameters as well as the extent of the shallow mantle provides the range of
40
ArxsD M depicted in ­ gure 5. Because of the large uncertainties in the shallow-mantle
concentrations, there cannot be a unique interpretation of the nature of the deep
mantle: 40 ArxsD M can be zero, corresponding to a closed system (undegassed?), but,
in the range of parameters, 40 ArxsD M can also correspond to a degassed or `overgassed’
deep mantle. In this framework, a degassed deep mantle is certainly as reasonable to
conceive as a primitive undegassed one. Moreover, the popping rock 4 He/40 Ar value
of 1.5 is similar to that of the closed-system (as discussed above), although we note
that the excess amount of 40 Ar in the shallow mantle could be very small, leaving
the deep mantle with an argon de­ cit close to ¡ 1:9 £ 1016 kg.
Even avoiding K/U uncertainties, the budget of argon in the various reservoirs
cannot be decoupled from that of potassium. This budget is probably not a strong
constraint on mantle models, and considerable improvement in our knowledge of
shallow-mantle concentrations is needed. In this framework, a hypothetical reservoir
the size of D 00 (ca. 2 £ 1023 kg), made of 2.5 billion-year-aged oceanic crust and containing 600 ppm of potassium, would provide a value of ¡ 1:2 £ 1016 kg for 40 ArxsD M .
This would imply an Arxs of ¡ 0:7 £ 1016 kg for the whole mantle except D 00 , which,
for example, could be explained by rather uniform K and 40 Ar concentrations of
50 ppm and 120 £ 10¡12 mol g¡1 .
6. A recycling model of the mantle
The isotopic and trace-element compositions in OIBs show that the sources of plumes
are often derived from sections of ancient oceanic crust. Every petrological unit
observed on the ocean ®oor has been identi­ ed: pelagic sediments (Blichert-Toft et
al . 1999a); plateaux (Gasperini et al . 2000); extrusive sections (sometimes altered)
and gabbros (Hauri & Hart 1993; Eiler et al . 1996; Chauvel & H´emond 2000). The
source of mantle plumes is then made from poorly mixed subducted oceanic plates.
Proposing that the `primitive’-like noble-gas signature is explained by recycling the
peridotitic part of convective downwellings is consistent with the presence of slab
material in plumes.
Segregation of the dense eclogitic oceanic crust at the base of the mantle is a
mechanism that could provide the recycled components in the plume source, poor
mixing in the boundary layer and deep chemical layering. The mineralogical assemblages in the basaltic crust at lower-mantle conditions have an excess density of a
few per cent relative to the bulk mantle (Kesson et al. 1994). Convection calculations show that segregation would produce an accumulation of eclogite at the base
of the mantle and a dense layer could grow with time according to a dynamic equilibrium between segregation due to gravity and plume entrainment (Christensen &
Hofmann 1994). The heat-producing elements are concentrated in the oceanic crust
and potentially destabilize the dense units. However, Christensen & Hofmann (1994)
have performed convection calculations where the radioactive elements are con­ ned
Phil. Trans. R. Soc. Lond. A (2002)
On the origin of noble gases in mantle plumes
2645
in the dense crust, corresponding to 5{20 times primitive mantle concentrations, and
they conclude that `the heating of the bottom pools from within does not play a
dominant role’. They only observed slight di¬erences in the amount of segregation.
Separation of the crustal component at the base of the mantle should satisfy mass
balance of trace elements, especially if sediments and plume products are considered
(Coltice & Ricard 1999). The size of the layer formed would then be comparable
to that of D 00 , which is known to be chemically distinct from the overlying mantle
(Wysession 1996). Like partial melting, segregation is a process of chemical di¬erentiation that sustains heterogeneity within the mantle. In the ­ rst investigation of
this model (Coltice & Ricard 1999), the partly degassed peridotite was considered
as a reservoir in order to perform box-model calculations. However, the recycling
model only requires that the source of plumes is less mixed than the shallow mantle.
The di¬erences between results from Coltice & Ricard (1999) and the present studies
can be mostly attributed to the limitations of box models in accounting for mixing
heterogeneities (Coltice et al. 2000).
7. Conclusions
The `primitive’-like noble gases observed in Hawaiian and Icelandic samples contain indications of whether or not there exists a geochemically primitive layer in the
deep mantle. If one exists, mass balance implies that the fraction of primitive material in these hotspots is very low and therefore, except for rare gases, incompatible
elements in OIBs and MORBs should be similar. It also implies that 3 He composition in the primitive mantle is richer than in carbonaceous chondrites. Furthermore,
the primitive-layer hypothesis requires that helium and neon should not have been
fractionated by extensive degassing of the shallow mantle, which is very unlikely.
On the contrary, partly degassed peridotite, observed in the shallow mantle, can
account for the source of `primitive’-like noble-gas OIBs, considering isotopic ratios
and 3 He/22 Ne data. The peridotite and recycled oceanic crust are ubiquitous components in the mantle, well mixed in the shallow part and poorly mixed in the deep
thermal boundary layer. These peridodites most likely contain some veins of primitive
material. In this model, `primitive’-like plumes would carry the same concentrations
of 3 He, 22 Ne or 36 Ar as the shallow mantle. No speci­ c assumptions on the relative
compatibility of helium and uranium have been used to reach these conclusions. The
mechanism for separating peridotitic and crustal components in the boundary layer
tapped by plumes can be the gravitational segregation of the denser oceanic crust at
the bottom of the mantle. This model would provide a deep layer of recycled components, corresponding to the extent of the D 00 layer. The chemical heterogeneity of
the mantle could then be a consequence of the combination of plate tectonics and
petrological density di¬erences.
References
Albarµ ede, F. 1998 Time-dependent models of U{Th{He and K{Ar evolution and the layering of
mantle convection. Chem. Geol. 145, 413{429.
Allµ egre, C. J. & Turcotte, D. L. 1986 Implications of a two component marble-cake mantle.
Nature 323, 123{127.
Allµ egre, C. J., Hofmann, A. W. & O’ Nions, R. K. 1996 The argon constraints on mantle structure.
Geophys. Res. Lett. 23, 3555{3557.
Phil. Trans. R. Soc. Lond. A (2002)
2646
N. Coltice and Y. Ricard
Anderson, D. L. 1998 A model to explain the various paradoxes associated with mantle noble
gas geochemistry. Proc. Natl Acad. Sci. USA 95, 9087{9092.
Ballentine, C. J. & Barfod, D. N. 2000 The origin of air-like noble gases in MORB and OIB.
Earth Planet. Sci. Lett. 180, 39{48.
Blichert-Toft, J., Frey, F. A. & Albarµ ede, F. 1999a Hf isotope evidence for pelagic sediments in
the source of Hawaiian basalts. Science 285, 879{882.
Blichert-Toft, J., Albarµ ede, F.& Kornprobst, J. 1999b Lu{Hf isotope systematics of garnet pyroxenites from Beni Bousera, Morocco: implications for basalt origin. Science 283, 1303{1306.
Bunge, H.-P., Richards, M. A., Lithgow-Bertelloni, C., Baumgardner, J. R., Grand, S. P. &
Romanowicz, B. A. 1998 Time scales and heterogeneous structure in geodynamic Earth models. Science 280, 91{95.
Chauvel, C. & H¶ emond, C. 2000 Melting of a complete section of recycled oceanic crust: trace element and Pb isotopic evidence from Iceland. Geochem. Geophys. Geosyst. 1, 1999GC000003.
Christensen, U. R. & Hofmann, A. W. 1994 Segregation of subducted oceanic crust in the
convecting mantle. J. Geophys. Res. 99, 19 867{19 884.
Coltice, N. & Ricard, Y. 1999 Geochemical observations and one layer mantle convection. Earth
Planet. Sci. Lett. 174, 125{137.
Coltice, N., Albarµ ede, F. & Gillet, Ph. 2000a 40 K{40 Ar constraints on recycling continental crust
into the mantle. Science 288, 845{847.
Coltice, N., Ferrachat, S. & Ricard, Y. 2000b Box modeling the chemical evolution of geophysical
systems: case study of the Earth’ s mantle. Geophys. Res. Lett. 27, 1579{1582.
Davies, G. F. 1981 Earth’ s neodymium budget and the structure and evolution of the mantle.
Nature 290, 208{213.
Dixon, E., Honda, M., McDougall, I., Campbell, I. & Sigurdsson, I. 2000 Preservation of nearsolar isotopic ratios in Icelandic basalts. Earth Planet. Sci. Lett. 180, 309{324.
Eiler, J. M., Farley, K. A., Valley, J. M., Hofmann, A. W. & Stolper, E. M. 1996 Oxygen isotope
constraints on the sources of Hawaiian volcanism. Earth Planet. Sci. Lett. 144, 453{468.
Eiler, J. M., Schiano, P., Kitchen, N. & Stolper, E. M. 2000 Oxygen-isotope evidence for recycled
crust in the sources of mid-ocean ridge basalts. Nature 403, 530{534.
Farley, K. A. & Poreda, R. J. 1993 Mantle neon and atmospheric contamination. Earth Planet.
Sci. Lett. 114, 325{339.
Farley, K., Maier-Reimer, E., Schlosser, P. & Broecker, W. S. 1995 Constraints on 3 He ° uxes
and deep-sea circulation from an oceanic general circulation model. J. Geophys. Res. 100,
3829{3839.
Ferrachat, S. & Ricard, Y. 2001 Mixing properties in the Earth’ s mantle: e® ects of the viscosity
strati¯cation and of oceanic crust segregation. Geochem. Geophys. Geosyst. 2, 2000GC000092.
Gasperini, D., Blichert-Toft, D., Bosh, A., Del Moro, P., Macera, P., T¶elouk, F. & Albarµ ede, F.
2000 Evidence from Sardinian basalt geochemistry for recycling plume heads into the Earth’ s
mantle. Nature 408, 701{704.
Grand, S. P. 1987 Tomographic inversion for shear velocity beneath the North American plate.
J. Geophys. Res. 92, 14 065{14 090.
Halliday, A. N., Lee, D. C, Tommasini, S., Davies, G. R., Paslick, C. R., Fitton, J. G. & James,
D. E. 1995 Incompatible trace elements in OIB and MORB and source enrichment in the
sub-oceanic mantle. Earth Planet. Sci. Lett. 133, 379{395.
Hanyu, T. & Kaneoka, I. 1997 The uniform and low 3 He/4 He ratios of HIMU basalts as evidence
for their origin as recycled materials. Nature 390, 273{276.
Harper Jnr, C. L. & Jacobsen, S. B. 1996 Noble gases and Earth’ s accretion. Science 273,
1814{1818.
Harrison, D. W., Burnard, P. G. & Turner, G. 1999 Noble gas behavior and composition in the
mantle: constraints from the Iceland plume. Earth Planet. Sci. Lett. 171, 199{207.
Phil. Trans. R. Soc. Lond. A (2002)
On the origin of noble gases in mantle plumes
2647
Hauri, E. E. & Hart, S. R. 1993 Re{Os isotope of HIMU and EMII OIBs from the South Paci¯c
Ocean. Earth Planet. Sci. Lett. 114, 426{436.
Hofmann, A. W. 1988 Chemical di® erentiation of the Earth: the relationship between mantle,
continental crust, and oceanic crust. Earth Planet. Sci. Lett. 90, 297{314.
Hofmann, A. W. 1997 Mantle geochemistry: the message from oceanic volcanism. Nature 385,
219{229.
Honda, M. & Patterson, D. B. 1999 Systematic elemental fractionation of mantle-derived helium,
neon, and argon in mid-oceanic ridge glasses. Geochim. Cosmochim. Acta 63, 2862{2874.
Jacobsen, S. B. & Wasserburg, G. J. 1979 The mean age of mantle and crustal reservoirs. J.
Geophys. Res. 84, 7411{7427.
Jean-Baptiste, P. 1992 Helium-3 distribution in the deep world ocean. In Isotopes of noble gases
as tracers in environmental studies. Proc. Consultants Meeting, International Atomic Energy
Agency, Vienna, pp. 219{240.
Jochum, K. P., Hofmann, A. W., Ito, E., Seufert, H. M. & White, W. M. 1983 K, U and Th in
mid-ocean ridge basalt glasses and heat production. Nature 306, 431{436.
Kellogg, L. H., Hager, B. H. & van der Hilst, R. D. 1999 Compositional strati¯cation in the
deep mantle. Science 283, 1881{1884.
Kesson, S. E., Fitz Gerald, J. D. & Shelley, J. M. G. 1994 Mineral chemistry and density of
subducted basaltic crust at lower mantle pressure. Nature 37, 767{769.
Kurz, M. D., Jenkins, W. J. & Hart, S. R. 1982 Helium isotopic systematics of oceanic island
and mantle heterogeneity. Nature 297, 43{47.
McDonough, W. F. & Sun, S. 1995 The composition of the Earth. Chem. Geol. 120, 223{253.
Mazor, E., Heymann, D. & Anders, E. 1970 Noble gases in carbonaceous chondrites. Geochim.
Cosmochim. Acta 34, 781{824.
Moreira, M. & Sarda, P. 2000 Noble gas constraints on degassing processes. Earth Planet. Sci.
Lett. 176, 375{386.
Moreira, M., Kunz, J. & Allµ egre, C. J. 1998 Rare gas systematics in popping rock: isotopic and
elemental compositions in the upper mantle. Science 279, 1178{1181.
Moreira, M., Breddam, K., Curtice, J. & Kurtz, M. D. 2001 Solar neon in the Icelandic mantle;
new evidence for an undegassed lower mantle. Earth Planet. Sci. Lett., 185, 15{23.
Pearson, D. G., Davies, G. R., Nixon, P. H., Greenwood, P. B. & Mattey, D. P. 1991 Oxygen
isotope evidence for the origin of pyroxenites in the Beni Bousera peridotite massif, North
Morocco: derivation from subducted oceanic lithosphere. Earth Planet. Sci. Lett. 102, 289{
301.
Pepin, R. O., Palma, R. L. & Schlutter, D. J. 2001 Noble gases in interplanetary dust particules.
II. Excess 3 He in cluster particules and modelling constraints on IDP exposures to cosmic ray
irradiation. Meteoritics Planet. Sci. 36, 1515{1534.
Porcelli, D. & Wasserburg, G. J. 1995 Mass transfer of helium, neon, argon, and xenon through
a steady-state upper mantle. Geochim. Cosmochim. Acta 59, 4921{4937.
Ricard, Y., Richards, M. A., Lithgow-Bertelloni, C. & LeStun® , Y. 1993 A geodynamic model
of mantle density heterogeneity. J. Geophys. Res. 98, 21 895{21 909.
Sarda, P., Staudacher, T. & Allµ egre, C. J. 1988 Neon isotopes in submarine basalts. Earth Planet.
Sci. Lett. 91, 73{88.
Sarda, P., Moreira, M. & Staudacher, T. 1999 Argon-lead isotopic correlation in mid-Atlantic
ridge basalts. Science 283, 666{668.
Sims, K. W. & De Paolo, D. J. 1997 Inferences about mantle magma sources from incompatible
element concentration ratios in oceanic basalts. Geochim. Cosmochim. Acta 61, 765{784.
Stuart, F. M. & Grenville, T. 1998 Mantle derived 40 Ar in mid-ocean ridge hydrothermal ° uids:
implications for the source of volatiles and mantle degassing rates. Chem. Geol. 147, 77{88.
Trielo® , M., Kunz, J., Clague, D. A., Harrison, D. & Allµ egre, C. J. 2000 The nature of pristine
noble gases in mantle plumes. Science 288, 1036{1038.
Phil. Trans. R. Soc. Lond. A (2002)
2648
N. Coltice and Y. Ricard
Turner, G. 1989 The outgassing history of the Earth’ s atmosphere. J. Geol. Soc. Lond. 146,
147{154.
Tyburczy, J. A., Frisch, B. & Arhens, T. J. 1986 Shock-induced volatile loss from a carbonaceous
chondrite: implications for planetary accretion. Earth Planet. Sci. Lett. 80, 201{207.
Valbracht, P. J., Staudacher, T., Malaho® , A. & Allµ egre, C. J. 1997 Noble gas systematics of
deep rift zone glasses from Loihi seamount, Hawaii. Earth Planet. Sci. Lett. 150, 399{411.
van der Hilst, R. D., Widiyantoro, S. & Engdahl, E. R. 1997 Evidence for deep mantle circulation
from global tomography. Nature 386, 578{584.
van Keken, P. E. & Ballentine, C. J. 1998 Whole-mantle versus layered convection and the role
of a high-viscosity lower mantle in terrestrial volatile evolution. Earth Planet. Sci. Lett. 156,
19{32.
Wysession, M. E. 1996 Continents of the core. Nature 381, 373{375.
Yatsevich, I. & Honda, M. 1997 Production of nucleogenic neon in the Earth from natural
radioactive decay. J. Geophys. Res. 102, 10 291{10 298.
Zindler, A. & Hart, S. 1986 Chemical geodynamics. A. Rev. Earth Planet. Sci. 14, 493{571.
Phil. Trans. R. Soc. Lond. A (2002)